Japan’s Computer Entertainment Supplier’s Association (CESA) has published a preliminary summary of its latest game-development research, ahead of the complete report scheduled for December. It includes one especially notable figure: 85.8% of Japanese game developers surveyed are now using generative AI in their work.
AI adoption has risen sharply
The results come from a survey of 1,349 developers carried out at the Japanese game developers conference CEDEC. That represents a substantial increase on last year’s report, which found that 51% of developers were using the technology.
Use extends beyond creating game assets. Developers also reported using generative AI tools such as ChatGPT and Copilot for broader administrative tasks. Among those using AI, 63% said they used it every day, while 22.8% used it occasionally. A further 8.6% said their companies had introduced guidance covering its use.
The role of AI in asset production has attracted particular scrutiny. A recent presentation from Level-5, the studio behind Dark Cloud and Professor Layton, prompted fans to question whether some displayed images had been generated by AI.
Developers report efficiency gains
Most respondents said generative AI had delivered some benefit. The commonly reported advantages were shorter production times and lower costs. Within the survey, 837 developers said AI had expanded creativity, while 738 said it had reduced technical barriers.
However, far fewer respondents connected the technology with a better player experience: only 147 said generative AI had benefited end users.
For more coverage of changing technology and practices across gaming, browse RETROCADE’s arcade news and industry articles.
Why the figures need context
The survey measures individual developers rather than companies. As a result, several respondents may work for the same studio and be subject to identical internal policies. It also does not show how widely AI is being used across each company.
CESA’s separate survey of member companies offers a more cautious picture. Seventy-four percent of those companies reported having policies governing generative AI. The results were otherwise almost evenly divided: 27 surveyed companies were not using generative AI, while 26 were using it in some form.
Only 12 companies reported using generative AI to create visual assets. Even fewer used it for story development, localisation, quality assurance or programming.
Policy is becoming as important as adoption
Generative AI use is clearly increasing within Japan’s game industry, but the development of formal rules may prove just as significant as the headline adoption rate. Many policies appear to have emerged in response to mixed player reactions and concerns raised online.
Support for these tools is stronger in East Asian markets including Japan and South Korea. Nevertheless, international criticism—including the backlash surrounding Crazy Taxi: World Tour—suggests game companies will need to consider how AI is implemented, not simply whether it can improve productivity or reduce production costs.
The same principle applies to arcade projects, where reliability and player-facing quality matter as much as efficiency. RETROCADE’s arcade machine buying guide explains the practical factors to consider when assessing hardware, while arcade parts and DIY arcade kits cover the hands-on side of building and maintaining a machine.
Leave a Reply